Remote Frog

COMMUNITIES ANALYTICS

Led enterprise-wide ONA initiative analyzing collaboration patterns across 1,500+ employees, uncovering hidden productivity opportunities:

• Built a communities model. Identified highest impact individuals and knowledge and collaboration brokers, enabling proactive succession planning to prevent high-risk transitions.

• Discovered isolated individuals and teams through graph analysis (Python, NetworkX) that allowed for targeted interventions increasing cross-team collaboration.

• Created automated pipeline (Meltano, ETL, Airflow, SQL, GraphX) processing 10K+ daily contributions.

• Influenced industry standards: Keynote presentations at HR Analytics Summit London & HR Analytics and AI Berlin 2025 (1000+ combined attendance), two peer-reviewed publications pending

Tech Stack: Elasticsearch, Meltano, Python (NetworkX, pandas, etc.), ELT (SQL, avdl, cucumber tests), ML clustering algorithms, Airflow

​CUSTOMER SUPPORT AND SATISFACTION

Rearchitected enterprise-wide customer support analytics ecosystem, bringing accuracy over 99.999%:

• Integrated disparate support systems (Zendesk Chat, Email, Social Networks, App ratings, and others) into a single source of truth, processing 80K+ weekly support async interactions and 13K+ sync chats.

• Data trust: Transformed a “we don’t trust data” culture into “what does the dashboard say?” mindset.

• Drove massive adoption: Implemented data literacy program and iterative dashboard design process resulting in 2,000+ weekly dashboard views.

• Technical Architecture: ETL pipeline (Meltano/Fivetran) → Data Lake (Spark/SQLT, avld, cucumber tests) → Self-service analytics (Looker/LookML) and integrated in other systems. API integrations feeding insights back to Slack, Zendesk, and alerting systems

ATTRITION PREDICTION MODEL

Impacted retention strategy by building a predictive attrition model incorporating organizational network analysis. The model combined measures like network centrality, and pagerank with clique communities belonging with traditional predictive factors

• Accuracy: 6x more accurate than random tests.

• Uncovered hidden risk: Discovered that belonging to a clique community is a factor missed by conventional models

• Massive ROI: Ability to avoid high-performer departures, saving in replacement costs and protecting from delayed project revenue

TALENT ACQUISITION REVAMPED

I went further than KPIs and led a portfolio of predictive talent acquisition studies that changed how to hire, measure success, and allocate recruiting resources:

• Quality of Hire: Developed QoH model linking pre-hire elements to 100-day performance, engagement, and culture fit signals to early alert HR of potential for intervention.

• Re-applicant goldmine: Uncovered that previously rejected candidates who reapplied were 2x+ more likely to become hired, leading to changes in hiring.

• Referral: Proved the exact added value of referrals in terms of hiring funnel behavior and likelihood of being hired.

• Analyzed rubric items in the hiring pipeline for some roles. We looked for correlations with success in the trial stage, which enabled some changes to the interviews scoring to maximize success.

• Technical Architecture: ETL pipeline (Meltano/Fivetran) → Data Lake (Spark/SQLT, avld, cucumber tests) → Self-service analytics (Looker/LookML) and integrated in other systems. API integrations feeding insights back to Slack, Zendesk, and alerting systems.

REMOTE LEARNING MODEL + SKILLS INTELLIGENCE

• Built skills intelligence system transforming static spreadsheet into dynamic analytics platform tracking skills across 300+ employees, enabling identification of critical capability gaps and measurement of training effectiveness.

• Developed data-driven remote training model through rigorous hypothesis testing and impact analysis, iterating based on participant outcomes. (ref)

• Core design principles: hands-on practical application, extended project-based learning, peer learning communities, and emphasis on information discovery and network-building skills.